Depending upon the jurisdiction, expert witnesses may be required to draft and sign an expert witness disclosure or expert witness declaration. The expert can find out what the disclosure or declaration needs to contain by asking retaining counsel.
Expert witness disclosures should be carefully drafted as these disclosures can be used against the expert witness in the case at hand or other future cases. In essence, what an expert writes in his expert witness declaration becomes a permanent part of the expert witness’s record. Retaining counsel often assist in the drafting process. Although there may be nothing wrong with such assistance per se, particular care should be taken to make sure that counsel does not ask you to put language in your declaration that you either do not agree with fully or do not fully understand.
If you are looking for a model or sample expert witness disclosure or declaration, you should be able to obtain these by asking retaining counsel for copies from past cases. An excellent resource for models and drafting advice is SEAK’s text; How to Write an Expert Witness Report. Please see below for an expert witness declaration/disclosure example excerpted from SEAK’s text on expert witness report writing:
Model Expert Witness Disclosure/Declaration
This report is most notable for:
- Easy to read.
- Use of examples to help explain opinion.
- Multiple reasons given to support conclusion that data in question was a trade secret.
- Objective language is used, even when refuting opposing expert’s declaration.
- No superfluous information included.
Areas of possible improvement:
Include materials reviewed.
Include list of authorities.
DECLARATION OF RICHARD A. KRYSTI
I. BACKGROUND AND EXPERIENCE
1. My name is Richard A. Krysti. I am the cofounder of Klein Marketing Services, a marketing consulting firm located in Wintersburg, Massachusetts. A complete copy of my curriculum vitae is attached as Exhibit 1.
2. I have held a variety of positions over the past 30 years in the field of marketing and marketing science. Many of these activities have involved the collection and analysis of data for marketing and sales-related purposes. Examples of these activities include:
- Analysis of geographic trends in consumer goods sales data, including cookies, crackers, snack foods, beer, autos, and energy.
- Analysis of trends in purchasing behavior of various demographic groups determined by relating individual item sales in supermarkets to the demographic makeup of each store’s trading area.
- Selection of representative markets for test marketing and experimentation based on U.S. Census data.
- Adjustment and weighting of consumer survey sample results to correspond with U.S. Census data.
- Projection of national and regional industry sales volumes and share based on sales data collected from supermarket scanners.
II. ASSIGNMENT AND SUMMARY OF OPINIONS
3. I have been asked by Huske and Allen (“Husk”) on behalf of ABC Mutual Automobile Insurance Company, ABC Fire and Casualty Company and ABC General Insurance Company (“ABC”) to provide this declaration in the above-captioned litigation between ABC and the California Department of insurance (“CDI”). As I understand the litigation, ABC has provided certain zip code level information, referred to as the Record A information, to the CDI as required by law. This zip code level information contains information on the type of insurance, the number of total exposures earned during the year , the number of new exposures acquired during the year, the number of exposures cancelled during the year, the number of exposures not renewed during the year, and the total earned premium for the year.
4. Husk has asked me to opine as to whether the zip code level information submitted to the CDI has significant value and whether firms would typically consider this sort of customer information to be a trade secret. My expertise is in the fields of marketing and marketing science, and I was asked to address this question generally based on my experience with companies in various industries, rather than as an issue somehow specific to insurance. I have analyzed markets for competitive purposes in many industries and the same knowledge, experience and expertise is applicable to each of the various industries. If asked to provide competitive advice and analysis to an insurance company, I would use the same knowledge, experience, and expertise that I have developed over 30 years of studying markets in numerous industries.
I apply that same knowledge, experience, and expertise here. Addressing the question with that understanding, in my opinion, the zip code level information contained in Record A would have significant value both to ABC and to competitors. Because of its strategic importance for internal decision‑making and its value to competitors, it is the sort of information firms tend to consider a trade secret and would take steps to prevent its discovery by competitors.
5. In my experience, zip code level information on customers and sales is widely used within companies and by competitors to determine patterns of behavior, identify market segments, develop and refine marketing and advertising strategies, and make other important marketing and business decisions. Indeed, the CDI’s own use of the Record A data to analyze penetration of regional markets demonstrates the importance and the value of using this zip code level data. In a similar manner, competitors can use this zip code level data to understand patterns of behavior and market segmentation, to target any vulnerabilities may have, and to avoid competing in areas where ABC is strong.
6. In the remainder of my declaration, I review some of the potential uses and sources of value of zip code level information such as the Record A data. These include:
- Using zip code level data for market share and market position analysis;
- Using zip code level data to develop regional advertising and marketing strategies;
- Using zip code level data to develop other targeted advertising and marketing strategies;
- Using zip code level data to assist in agent recruiting efforts;
- Using zip code level data to evaluate the success of advertising efforts; and
- Using zip code level data to measure the impact of price changes.
7. I conclude by commenting on some of the arguments being raised that seek to deny the value and trade secret status of the Record A data. These include:
- Arguments that the data are not detailed enough to have any value;
- Arguments that Record A data provide no additional value since rate data is already available at the zip code level; and
- Arguments that if the data were released for all companies, all companies would be on equal footing.
III. SOURCES OF VALUE FROM ZIP CODE LEVEL DATA
8. As I understand it, earned premium and earned exposure data are currently available in the annual statements filed by ABC and other insurers at the state‑wide level for California. The relevant question is what additional value may result from the data being available at the zip code level, and whether this value may be great enough that the data should be considered to be a trade secret. In my experience, zip code level information such as this is commonly used for a wide variety of marketing and other business purposes. It is also information that many companies consider confidential and would not want to fall into competitors’ hands. In this section, I will discuss several of these uses in more detail.
A. Value of the Record A Data for Market Share and Relative Market Position Analysis
9. One of the most important uses of ABC’s Record A data would be as part of either a market share analysis or a relative market position analysis at the zip code level. Market share analysis is the comparison of the share of the market of one company relative to all of its competitors and relative to the market as a whole. Relative market position involves a comparison of one company’s position relative to another. Both of these are key elements of market intelligence and are useful for measuring and evaluating the performance of a product or service.
10. Currently market share information is available annually by line of insurance for the entire state of California. However, given the tremendous size (over 34 million people) and geographic diversity of the population in California, this statewide market share provides only a limited picture of the market and is much less information than a company would want to have in order to understand a competitor’s actions.
11. Were ABC’s Record A data to be available, other companies would be able to develop relative market position analyses at the zip code level by line of insurance, comparing their business to ABC. This type of analysis is of particular importance when the lines of personal insurance, including automobile and homeowner’s insurance. Thus while most carriers would perhaps not care very much about how they were performing relative to a small insurer, they are likely to care very much about how they were performing relative to the largest company in the market.
12. Manufacturers and marketers routinely spend hundreds of thousands of dollars to buy market share and competitor market information. Indeed, there are a number of companies that specialize in collecting and selling these data, including Information Resources, A. C. Nielsen, NPD, and others. In each case, however, the information is collected from a sample of stores, customers, etc. rather than a complete enumeration of all the sales of all the competitors in all the outlets. The Record A data at the zip code level are much more precise, and thus even more valuable, than the best data available in most other industries. The Record A data represents a census, and the resulting market shares can be calculated exactly with zero sampling error.
13. If a competitor wanted to approximate the information available in the Record A data, it would be necessary to interview respondents in every zip code and ask them details about each type of insurance they own. Due to normal sampling variation, a very large number of respondents would be needed to provide usable measures of market share at the zip code level. For example, 400 respondents would be required from each of the over 1,800 zip codes in California to be able to estimate the market share in a zip code plus or minus 4% (20% plus or minus 4% = 16% to 24%). The cost of such a data collection effort would exceed $20 million per year and yield information that is significantly inferior to the Record A data due to factors such as normal sampling error, respondent bias due to potential language issues, and the reluctance of respondents to answer questions about potentially sensitive insurance-related issues.
14. I have reviewed the declaration of J. Robert Hult, dated May 1, 2000. In his declaration, he states that the insurer, “knows where it is gaining or losing market share to other competitors because agents and others selling the insurance know such information through their on-the-ground sales activities.” Based on my experience in marketing and analysis of marketing data, such on-the-ground, anecdotal information is often inaccurate and inadequate for use in marketing activities. While there may be some value to such information, it is certainly not a substitute for the sort of reliable, detailed information found in the Record A data.
B. Value of the Record A Data for Developing Regional Marketing and Advertising Programs
15. In addition to being valuable simply as market intelligence, the zip code level information available in ABC’s Record A data has additional uses that are of equal or even greater value. For example, using Record A data, competitors would be able to analyze their position relative to ABC’s and develop marketing and advertising strategies to exploit this information in distinct local markets. ABC is obviously a formidable competitor, but the zip code level information available in the Record A data can be used to determine if there are any weaknesses that might be exploited at the regional level.
16. As an example of how this might be done, consider the following example. If a competitor insurance company sought to target ABC, it might begin by using the Record A data of ABC and its own Record A data to identify those zip codes or groups of zip codes where it had experienced some success over time at the expense of ABC. It could then use this information to target its media advertising in those areas where it had already shown that is was most likely to attract customers away from ABC.
17. California consists of twelve distinct broadcast media markets ranging in size from Los Angeles (number 2, nationally) to Eureka (number 189). By focusing advertising dollars on those media markets where they would have the greatest impact, a competitor could spend hundreds of thousands or even millions of dollars less than it might otherwise have to in order to obtain a given number of former ABC customers from a major advertising effort.
18. Based on my review of the reports of the insurance industry over the past few years, I do not believe that this is simply a theoretical possibility. The information contained in Record A data would be of particular use to a company specifically targeting ABC. An example of such a company would be BestBuy Insurance Company. The CEO of BestBuy Insurance has publicly stated that the company’s goal is to surpass ABC in auto insurance premiums written by 2010. He has further stated that his company is “using specifically designed strategies to gain market share.”3 The Record A data provides the type of information BestBuy could use to both target its efforts and evaluate the success of those efforts. With this information BestBuy could precisely target the exact zip codes where ABC is most vulnerable
19. One strategy being used by BestBuy is to “expand its direct business sales via direct mail, the Internet, the telephone and to affinity groups. BestBuy claims that its direct business comes from taking business away from the captive agents of its competitors. These are the type of agents used by ABC, and the Record A data would be extremely valuable in determining the areas where this direct marketing effort should be focused. More evidence of this tightly focused strategy is the use of television commercials “specifically designed for rollout on local cable outlets.”5 By using the Record A data, BestBuy could choose the specific media alternatives and specific areas that would best meet their stated needs.
C. Value of the Record A Data for Developing Other Targeted Marketing and Advertising Programs
20. Zip code level data can also be combined with other data to develop targeted marketing and advertising programs. The U.S. Census Bureau publishes detailed information on age, income, employment, ethnicity, family size, education, housing characteristics, pets, and other information. This is collected by the Census Bureau at what is known as the “block group/enumeration district” level, an even finer level of geographic detail than zip code. Independent companies will aggregate this data for a fee to match any geography of interest, including zip code.
21. Analyzing market share and changes in market share in relation to the demographic make up of the customers in a zip code or group of zip codes allows a company to see where and why its competitors succeed and fail with different customer groups. This can provide critical competitive intelligence concerning their marketing strategy and the relative success of that strategy. It can also allow competitors to design strategies to target those population groups where ABC or other competitors may be vulnerable, as identified by the patterns of customers entering and leaving ABC’s customer base. This would be particularly true if the Record A data were released for a number of years, thus allowing competitors to identify trends of customers entering and leaving ABC’s customer base in each zip code and each line of insurance.
22. As an example of how this might be done, consider the following. If a competitor insurance company sought to target ABC’s most vulnerable customer groups, it could use the Record A data spanning several years and combine this with the Census Bureau data. The results of this analysis might show that ABC was losing customers in the markets with higher concentrations of young families. This information could then be used to develop direct mail marketing and advertising campaigns specifically focused towards young families. By targeting only those groups with the largest potential impact, a competitor could use its marketing resources to the maximum advantage.
D. Value of the Record A Data for Agent Recruiting Purposes
23. Another possible use of the Record A data, and thus another source of its trade secret value, is as a tool for recruiting new sales agents. ABC, the largest insurer in California, has offices throughout the state of California staffed with agents under exclusive agency contracts. Inevitably, some of these agents and their employees are going to be more successful than others. By using the Record A data strategically, competitors could significantly improve their own agency force at the expense of ABC by raiding ABC’s best offices.
24. Currently information on where ABC’s agents are located is available from the yellow pages or other public sources, and is not considered by ABC to be a trade secret. This information includes the zip codes of these agents and the agent names. Combining this information with ABC’s and a company’s own Record A data would provide opportunities to improve a company’s sales force through the targeted recruiting of ABC agents.
25. For example, a competitor could use the Record A data to identify zip codes where ABC was experiencing the greatest sales growth. It could then attempt to recruit either the primary ABC agents or the employees of those agents in those zip codes. While the data do not provide details as to exactly which agent or employee of the agent would be the cause of the success, this sort of information would nonetheless be of value in any effort to recruit new agents.
E. Value of the Record A Data for Advertising Evaluation and Strategy
26. If zip code level data such as the Record A data were available for all competitors, it could also be used by a competitor in the measurement and evaluation of advertising media. Insurance companies such as ABC currently spend tens of millions of dollars a year on a wide variety of media in order to promote brand awareness and deliver their unique marketing messages.
27. Because the Record A data measures market share and customer activities exactly and without error, evaluation of advertising is feasible even if the effect is small relative to the other factors affecting market share. The advertising weight of each company will naturally vary because every television station, radio station, and newspaper has a different pattern of coverage. Each of these media alternatives reports its reach and penetration by geography, and syndicated information services track advertising spending by company and media outlet. Thus it is possible to calculate the broadcast advertising weight each company is applying against each geographic area. This information can then be used to determine the impact of advertising on the market share achieved and to evaluate the payoff for increases or decreases in advertising. It will also indicate which geographic areas are most responsive to different types of advertising so that these funds can be allocated in the most efficient way possible.
28. Any properly advised company wishing to target ABC would see this data as being uniquely valuable. This information can be particularly important for companies that are aggressively seeking to grow their insurance business in California. Because ABC is the largest insurer in California in the major lines of personal insurance, it would be the obvious target for smaller companies wishing to grow their business. By using the Record A data to determine where advertising is most effective, competitors can target their advertising efforts to maximum advantage.
29. In addition to its value in evaluating the success of overall advertising efforts, the effectiveness of particular advertising themes can also be identified using zip code level information. Companies spend a great deal of time and effort developing advertising messages that “position” the company and its products in the minds of potential customers. Particular advertising themes are well established, but others are much more short-lived or may be focused on a particular demographic group or geographic area. When campaigns do change, an evaluation of this impact on the market is almost impossible without very precise market share information. By analyzing the year-to-year changes in market share coinciding with changes in advertising message, the overall effectiveness of the message can be measured. In addition to measuring the effectiveness of one’s own advertising messages in increasing market share, the Record A data could also be used to measure of the effect on retention, turnover, and market shares of advertising efforts by competitors.
30. The Record A data, the advertising weight data, and the Census Bureau data previously discussed can be combined to analyze the demographic makeup of the zip codes with the largest and smallest response to advertising messages. These changes will allow competitors to develop their advertising to counter competitive gains or exploit competitive losses. This analysis is only possible because the Record A data is a complete census, not a sample. Market share information derived from surveys is typically too imprecise due to sampling error to be used for this purpose.
F. Value of the Record A Data for Measuring the Impact of Price Changes
31. Measurement of price elasticity—the change in business resulting from a change in price is also possible by using the Record A data in conjunction with publicly available pricing data. Here the result is not only the loss of competitive advantage to ABC, but the potential for some consumers to see the cost of their insurance increase. The prices insurance companies charge are based on cost and risk, which can vary by location. Every insurance company is likely to have a unique cost structure and risk experience in each location where it operates. Thus prices can vary by zip code.
32. The result of all of this is that a customer in a particular location may be exposed to different prices from each company, with prices typically changing every year. When these year-to-year price changes are correlated with the changes in market share measured by the Record A data, the relative price sensitivity of the customers in every zip code can be measured for each insurance company. These measured price elasticities can be used to determine the potential profit impact of any price changes.
33. The following example illustrates how this might work. A competitor could use the publicly available insurance rate information along with the zip code level Record A data to identify zip codes or demographic groups that are relatively more or less sensitive to price changes. Assuming it was profitable to do so, the insurer could then undercut ABC’s prices in the markets with the largest potential benefit and maintain or possibly even raise the rates in those zip codes where price changes appeared to have little effect. This is an analogous strategy to the price discrimination strategy used by the airlines to identify business travelers and charge them more.
34. In summary, the nature of the zip code level Record A data makes it particularly valuable as a competitive tool. There is no feasible or affordable means available to competitors to develop comparable information using surveys or other means. Because this information is valuable and has multiple uses, it is logical and reasonable that an insurance company with a substantial presence would consider this information to be a trade secret and do everything possible to prevent it from falling into a competitor or future competitor’s hands. This is particularly true for ABC, which is the largest insurer in California, and is therefore an obvious target for many of its competitors.
IV. ANALYSIS OF STATEMENTS CLAIMING THE RECORD A DATA IS NOT A TRADE SECRET
35. It is my understanding that various arguments have been raised in this litigation concerning the issue of the value of the Record A data, and thus by extension its status as a trade secret. In particular, I have read the Declaration of Andrew Hult, FCAS, MAAA, In Support of Intervenors’ Motion For Summary Judgement, dated May 1, 2000 and filed July 28, 2000.
In this section, I will comment on certain of these arguments.
A. The Data Is Not Detailed Enough to Have Any Value
36. According to this argument, zip code level data has little or no value since the names, addresses, and telephone numbers of policyholders are not included. While it is true as a general rule that the more detailed the data, the more value it provides, it certainly does not follow that anything less than a complete list of customers and their history has no value. The existence of much more important data does not bear on whether a particular class of information has value.
37. As discussed earlier, analysis of zip-code level aggregate sales information is routinely performed in the marketing and market research departments of most large companies. Specialized computer software to facilitate such analyses has been in wide use for over 20 years, and companies pay large sums of money to obtain and analyze zip code level data.
38. A variant of this argument states that only if actual loss or profitability data were available at the zip code level would the Record A data actually have value. Again, while it is true that having more detailed data or additional fields would increase the value of the Record A data, it is still very valuable information even without these additional fields.
B. The Record A Data is of No Value Since the Premium Rate Data is Already Publicly Available
39. Insurance companies such as ABC currently file their rate information at the zip code level, and thus these data are already available to competitors and the general public. The rate data provides information on the price per unit of insurance but does not provide the zip code level data set forth in Record A.
40. Undoubtedly, price is an important driver of customer choice, but if it were the only factor, the company with the lowest rates would have all the business. Consumers choose many things on the basis of more than just price. Issues of trust, brand loyalty, reputation, customer service, company and brand image, advertising, personal interaction with the agent, prior positive or negative experience with the company, word-of-mouth recommendation of friends and neighbors, and other factors all play a role in the consumer decision process.
41. Zip code level market data such as the Record A data can be used to develop models to estimate the importance of these non‑price factors in different areas. This can then provide strategic intelligence for competitors for targeting ABC’s customers for their new business. Any firm that does not consider customer data such as the Record A data as trade secrets puts itself at risk in the marketplace. While “reverse engineering” may not be an actuarial concept, it is a common marketing science notion, and is used in practice by many large, sophisticated companies.
42. A variant of this argument has been made that the anecdotal “on-the-ground intelligence” of agents could provide most, if not all, of the value provided by the Record A data. While this on-the-ground intelligence may have some value, in a state as large and as important as California, doing any systematic analysis of markets would require the sort of information that is included in the Record A data. To consider that anecdotal impressions of the market can provide anything approaching the precision of the Record A data is significantly overstated.
43. In fact, many of the studies described in this declaration are only enabled by the precision of the Record A data. These same studies would not be nearly as accurate—and therefore as valuable—if the only available data was anecdotal feedback from agents. Among other problems, differences in the perception of agents as to the changes in the market would result in inconsistent data across regions, thus resulting in biased data and analysis.
C. If the Data Were Released for All Companies, All Companies Would be on Equal Footing
44. This argument states that if the data is released for all companies and all years, then it would no longer be a trade secret for anyone. This argument has two problems. First, it amounts to an argument that says that by destroying everyone’s trade secrets, everything will be fine. This does not change the fact that by releasing the information, the trade secret is being damaged or destroyed.
45. The second problem with the argument is that it fails to recognize that the value of the information can vary among companies, and that the release of such information could harm some companies more than others. For example, since ABC is the largest insurer, it would be a logical insurer for many smaller companies to target for sales growth. In those circumstances, the small companies would derive tremendous value from ABC’s information, which causes ABC to place a high value on its confidentiality. It is unlikely to be true that ABC would derive the same amount of value from obtaining the Record A data from most smaller insurers.
46. Given this potentially large difference in the valuation of the Record A data, it is not surprising that ABC might have a stronger interest than some smaller companies in protecting the Record A data. It is likely that several smaller companies would actually benefit from across the board release because they would obtain access to ABC’s Record A data and the Record A data of other historically successful insurers, which is of greater value than the confidentiality of their own data. To ABC and other historically successful insurers, on the other hand, the confidentiality of their data is likely to have more value than does access to other insurers’ data.
47. It is certainly true that ABC likely has data in its possession that it considers even more valuable than the Record A data. Zip code level data on loss experience, specific customer behavior, specific agent sales, and other data would be extraordinarily valuable to competitors. However, it is also the case that the zip code data contained in Record A would also be very valuable to competitors, and would typically be considered to be a trade secret.
Using modern marketing technology, this sort of data can prove extremely useful to competitors. As an expert in the use of data in marketing, I can attest to the fact that this sort of information is both very useful and very valuable, and should thus be considered to be a trade secret when claimed by the firm and treated as confidential.
I declare under penalty of perjury under the laws of the State of California that the foregoing is a true and correct statement of my opinions and the supporting facts and that this declaration was executed on August 9, 2000 at Wintersburg, Massachusetts.